摘要
美国海军运用神经网络模式识别技术,开发了基于多传感器多数据融合的船用早期火灾探测系统(EWFD),试验证明该系统总体性能优于商业感烟探测系统。之后,为寻求一种不依赖火灾烟气羽流自然扩散、能够监视整个空间、且具有较强抗误报能力的新型探测技术,又研发了一种以视频探测技术为主,辅以光谱、声学探测的非接触型立体传感探测(VS)原理样机。试验验证了该系统的探测能力、抗误报能力、响应速度等性能优于商业探测系统,并且能有效识别管道破损和水害事件。
An early warning, muhi-criteria, fire detection system, named EWFD system was first developed by USA navy and demonstrated to provide reliable warning of actual fire conditions in less time with fewer nuisance alarms than smoke detection systems. After this work, various spectral and acoustic signatures, new video imaging techniques, and image recognition methods had been investigated and integrated into a muhi-sensory prototype system known as the volume sensor to develop new detection capabilities for improved damage control and real-time situational awareness on USA navy ships. The prototype systems were shown to nutperform the commercial systems for flaming and smoldering fires with a high immunity to nuisance sources. In addition, the prototypes accurately identified pipe ruptures and flooding events.
出处
《船海工程》
2013年第4期71-76,80,共7页
Ship & Ocean Engineering
